semantic-kernel on promptflow

[mylee04]

Hello, I'm trying to figure out how to utilize semantic kernel on promptflow. Do you have any tutorial for this? for example,

kernel = sk.Kernel()
kernel.add_chat_service(
    "chat_completion",
     sk.AzureChatCompletion(
         deployment_name=AZURE_OPEN_AI_CHAT_COMPLETION_DEPLOYMENT_NAME,
         endpoint=AZURE_OPEN_AI_ENDPOINT,
         api_key=AZURE_OPEN_AI_API_KEY,     ),)

here, I don't need to use AzureChatCompletion part in promptflow, they have all these things if I make a llm node,

[harbhajan21]

You can just store the credentials of Azure Open AI in .env file and later you can retrieve it. Here, I'm attaching the YouTube video link for the reference.
Link: https://www.youtube.com/watch?v=90hhJHTWz50&t=13s

[mylee04]

Thank you, and have you ever tried it on promptflow?

[harbhajan21]

Yes, I tried and worked on promptflow.

[mylee04]

Hmm, okay, I will try that too. But I didn't solve it yet.